Scribd
Upload
33 views4 pages

Individual Assignment 3 Guideline

This document provides instructions for an assignment on applying cross validation and early stopping techniques in machine learning models. The objective is to familiarize students with K-fold cross validation and early stopping. Students will use a mobile phone price classification dataset to classify prices as low or high using KNN and neural network models. For the first part, students will perform K-fold cross validation on a KNN model to evaluate performance and select optimal k value. For the second part, students will implement early stopping on a neural network to prevent overfitting during training. The assignment requires delivering results like optimal k value, score lists, model performance discussion, training loss plots and test accuracy values.

Uploaded by

Abner ogega
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
33 views4 pages

Individual Assignment 3 Guideline

This document provides instructions for an assignment on applying cross validation and early stopping techniques in machine learning models. The objective is to familiarize students with K-fold cross validation and early stopping. Students will use a mobile phone price classification dataset to classify prices as low or high using KNN and neural network models. For the first part, students will perform K-fold cross validation on a KNN model to evaluate performance and select optimal k value. For the second part, students will implement early stopping on a neural network to prevent overfitting during training. The assignment requires delivering results like optimal k value, score lists, model performance discussion, training loss plots and test accuracy values.

Uploaded by

Abner ogega
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 4

BUSI 651 – Individual Assignment 2

Cross Validation and Early Stopping in Mobile Price


Classification

Note: Do not include the questions as well as dataset in your submission (to avoid similarity with
other submissions)

Objective:
The objective of this assignment is to familiarize you with two essential concepts in machine
learning: K-Fold Cross Validation and Early Stopping. You will apply these techniques to a
classification task using a mobile price classification dataset. Through this assignment, you will
learn how to implement K-Fold Cross Validation to assess model performance and how to use
Early Stopping to prevent overfitting during model training.

For both parts of this assignment, you'll use the provided 'mobile.csv' dataset to classify mobile
prices as low (0) or high (1).

Part 1: Cross Validation and KNN Classification


In this section, you will learn about k-fold cross validation and its application to the K-Nearest
Neighbors (KNN) classification problem.

Step-by-Step Procedure:
The step-by-step procedures (11 steps) to apply cross validation and model performance
evaluation is provided in the file ‘Classification Metrics and K-fold CV.ipynb’ which can be find
under the Additional Materials folder.
IMPORTAN NOTE: some parts of the codes are missing and denoted by ‘???’ and left to you to
complete. This is an easy part provided that you follow the instructions carefully. (Also read the
green comments before the codes which help you fill the ??? sections easier)
Deliverables:
1. At the end, provide a short summary of the following results: optimal k (based on ‘accuracy’,
‘f1_score’), ‘accuracy’ and ‘f1_score’ score lists (20 for 20 different k’s) according to k-fold cross
validation.

Optimal K

Evaluation Metric:
accuracy

Evaluation Metric:
f1_score

Score List

Evaluation Metric:
accuracy

Evaluation Metric:
f1_score

2. The screen shots of confusion matrix for steps 10, 11.

3. A short discussion of whether the model is overfitted, underfitted or proper fit? Interpret the
results to understand the model's performance and potential bias-variance trade-offs.

Discussion on
model fitness MAXIMUM 5 Lines
(question 3)
Part 2: Early Stopping and Artificial Neural Network Classification
In this section, you will learn about the concept of Early Stopping and how to implement it to
prevent overfitting during model training.

Step-by-Step Procedure:
The step-by-step procedures (9 steps) to apply cross validation and model performance
evaluation is provided in the file ‘Early Stopping.ipynb’ which can be find under the Additional
Materials folder.
IMPORTAN NOTE: some parts of the codes are missing and denoted by ‘???’ and left to you to
complete. This is an easy part provided that you follow the instructions carefully. (Also read the
green comments before the codes which help you fill the ??? sections easier)

Deliverables:
1. Screen shot of training and validation loss over epochs.
2. The train and test accuracy.
3. Do you think the model gets overfitted after 100 epochs? If no, provide your reasons, and if
yes, what is the optimal number of epochs to stop training at?

accuracy

Test

Train

Discussion on
model overfitting MAXIMUM 5 Lines
(question 3)
IMPORTANT: The report submission should include ONLY the following components:
Part 1:
• Three Tables: Optimal K, Score list, and discussion on model fitness.
• Two Screenshots: Train and test confusion matrix.

Part 2:
• Two Tables: Train and test accuracy, and discussion on model overfitting.
• One Screenshots: training and validation loss over epochs.

Please adhere to this specified structure and format for your submission.

1.0 Grading Rubric


Key Points Grade Allocation
(%)
Format (font type, size, table, formulas), overall 10
content, including references if required (APA
Style)
Results, analysis and assumptions (parts 1-4) 70
Novelty and creativity in solution (part 5) 20

N.B. Failure to comply with the above would result in low grades.

2.0 Format and Deadline


Submission: a single word/pdf document

Deadline: As posted in Course Shell

You might also like